Data Intermediary Networks
(New co-authored paper)
Kelsie Nabben
August, 2024
TLDR:
This post shares a new co-authored paper that shifts the emphasis from ‘data DAOs’ to networks of data intermediaries via data mesh architecture. I also discuss some of the thinking and context behind the research, and other milestones in the project, that were generously supported by a Filecoin Foundation developer grant (‘Improving Socio-technical Resilience in IPFS: Part 2’).
Brief Abstract of New Paper
A data mesh is about distributed data architecture at scale. Our paper argues that “Data Intermediary Networks”, composed of data products built on data mesh infrastructure, can foster co-operation and co-regulation among data product operators, data providers and data consumers. We argue that this leads to an increase in the variety of economically viable data products, along with an expansion of new markets within which those products can be networked together into meshes with the capacity to generate new kinds of value. We then outline the technical, operational, and governance architectures that underpin these networked data economies.
See:
Sisson, David; Nabben, Kelsie; Ben-Meir, Ilan; and Zargham, Michael. (2024).“Data Mesh Architecture: Interoperability, Co-Operation, and Co-Regulation” SSRN. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4880709.
According to Dietrich Ayala, long time IPFS project contributor, these concepts push on some of the intractable conundrums in the IPFS world. In particular, who “owns” content, and how to trace the data lifecycle of who publishes, pays, uses, and cares about it. These questions, that are straightforward in Web2, are incredibly difficult to answer in IPFS and other Web3 architectures. “This intermediary/mesh work helps us understand the shape of tools to meet user needs in p2p-available environments”, states Ayala.
Further Context…
I have been thinking about Decentralised Autonomous Organisations (DAOs) as data architecture for a while.
Part 1 of my research on improving resilience in IPFS culminated in a research paper and chapter of my PhD that looked at how social arrangements impact resilience in peer-to-peer networks (see: Nabben, K. (2022). “Decentralized Technology in Practice: Social and technical resilience in IPFS”).
Some of this began with writing about how DAOs may evolve existing governance architectures for data, such as DAOs as Data Trusts. I then moved to developing practical tools to help designers and developers think through social arrangements surrounding data management to improve resilience, such as the Decentralised Data Governance Pattern Library.
This most recent paper began as an attempt to analyse and operationalise the concept of ‘data DAOs’. Through desktop research and interviews, the output evolved beyond the rather nascent and mostly theoretical concept of data DAOs to explore the more applicable concept of ‘data meshes’. It was undertaken in conjunction with domain experts from BlockScience.
What guided the research was the question of ‘how is value derived from data?’. This led to an exploration of data as a product and data processing pipelines. A linchpin of this work was the use of an Actor Network Theory approach to identify data intermediaries as key actants in the creation of value in data economies. What’s interesting about this is that a focus on intermediaries seems counterintuitive in a domain of peer-to-peer architectures and ‘disintermediation’. Yet, these actors are critical in the production of value from data, and how they relate to one another determines whether a data ecosystem is centralised or decentralised.
Combined with the technical expertise of my colleagues, this led us to more deeply explore the data mesh architecture and develop the concept presented in the paper of ‘data intermediary networks’. I also plan to share a forthcoming academic paper on data intermediaries from an ethnographic case study perspective, which was presented on a panel for the ‘Platform Economies Research Network’ (PERN) conference in New York in April, 2024.
Enjoy the read and feel free to be in touch with questions, comments, and ideas.
Kelsie.